⚡ electrical engineering

Near-Field Localization via Reconfigurable Antennas

This paper proposes a joint baseband and electromagnetic precoder design for reconfigurable antennas that synthesizes specific beampatterns to significantly minimize user equipment positioning error bounds in near-field scenarios, outperforming traditional non-reconfigurable arrays.

Original authors: Alireza Fadakar, Yuchen Zhang, Hui Chen, Musa Furkan Keskin, Henk Wymeersch, Andreas F. Molisch

Published 2026-03-18

📖 4 min read☕ Coffee break read

CC BY 4.0

Original authors: Alireza Fadakar, Yuchen Zhang, Hui Chen, Musa Furkan Keskin, Henk Wymeersch, Andreas F. Molisch

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine you are trying to find a friend who is lost in a large, dark warehouse. You have a flashlight, but it's a very special kind of flashlight.

The Problem: The "Fixed Beam" Flashlight

Most traditional wireless systems (like the cell towers in your phone) use fixed-beam antennas. Think of these like a standard flashlight with a lens that can't change. You can point the flashlight in a specific direction, but the shape of the light beam is always the same.

The Limitation: If your friend is standing right next to the wall, a wide beam wastes light on the empty space. If they are far away, a narrow beam might miss them if you aren't pointing exactly right. In the "near field" (when the object is close to the antenna), the light waves curve like ripples in a pond, making it even harder to pinpoint exactly where the friend is using a simple, fixed beam.

The Solution: The "Shape-Shifting" Flashlight

This paper introduces Reconfigurable Antennas (RAs). Imagine if your flashlight could instantly change the shape of its beam.

Sometimes, you could make the beam wide and flat to scan a whole area quickly.
Other times, you could make it a tight, focused laser to hit a tiny spot.
Or, you could shape it like a donut to avoid a specific obstacle.

The researchers call this "synthesis." Instead of just pointing a fixed beam, the antenna actively reshapes its light (radio waves) to fit the situation perfectly.

The Goal: Finding the "Lost Friend" (Localization)

The paper isn't about sending more data (like streaming a movie); it's about localization—figuring out exactly where a device (the User Equipment or "UE") is located.

The Challenge: In the "Near Field" (close range), the waves are curved. Old methods assume the waves are flat (like sunlight hitting the earth), which causes errors when the object is close.
The Innovation: The authors created a smart system that uses these shape-shifting antennas to "feel" the curvature of the waves and calculate the exact 3D position of the device.

How They Did It: The "Tuning" Process

The researchers developed a two-part strategy, like a conductor leading an orchestra:

The Digital Conductor (Baseband Precoder): This decides when to send the signal and how much power to use.
The EM Conductor (Reconfigurable Precoder): This decides what shape the antenna beam should be for that specific moment.

They solved a complex math puzzle to find the perfect combination of "when to send" and "what shape to make" so that the signal bounces back with the most information possible. They used a mathematical tool called the Fisher Information Matrix (think of it as a "clarity score") to ensure their design gave the sharpest possible picture of the user's location.

The "Guessing Game" Strategy

Since the system doesn't know exactly where the user is at the start (otherwise, why are we trying to find them?), they use a clever two-step approach:

The Rough Sketch (Coarse Stage): They send out a few different beam shapes to scan a wide area, like casting a wide net to see where the fish might be. This gives a "good enough" guess.
The Zoom In (Refinement Stage): Once they have a rough idea, they switch to a high-precision mode, using the shape-shifting antennas to focus intensely on that small area to get the exact coordinates.

The Results: Why It Matters

The paper ran simulations to test this idea. The results were impressive:

Better Accuracy: The shape-shifting antennas found the user much more accurately than traditional fixed antennas, especially when the user was close (Near Field).
Robustness: Even when there was "noise" or interference (like other people shouting in the warehouse), the new system held its ground better than the old ones.
Efficiency: They proved that by using a specific set of mathematical "building blocks" (called Spherical Harmonics) to create these beam shapes, they could get the best results with the least amount of computing power.

The Bottom Line

This paper is about giving wireless antennas a new superpower: the ability to change their shape on the fly. By doing this, we can locate devices with incredible precision, even when they are right next to the antenna. This is a huge step forward for technologies like Augmented Reality (AR), autonomous robots, and the Internet of Things (IoT), where knowing exactly where something is, down to the centimeter, is critical.

1. Problem Statement

The paper addresses the challenge of near-field (NF) localization for User Equipment (UE) in wireless communication systems.

Context: While localization is critical for Integrated Sensing and Communication (ISAC), most existing research focuses on far-field (FF) scenarios where wavefronts are planar. In the near-field, wavefronts are spherical, requiring different algorithms that account for both angle and distance.
Limitation of Current Tech: Conventional antenna arrays use elements with fixed radiation patterns. This creates a trade-off: narrow patterns offer high gain but limited coverage, while broad patterns offer coverage but lower peak gain. Fixed patterns cannot dynamically adapt to optimize for specific localization tasks.
Gap: There is a lack of research applying Reconfigurable Antennas (RAs) to NF localization. RAs can dynamically reshape radiation patterns, offering a new degree of freedom (DoF) in the electromagnetic (EM) domain, but their potential for sensing has been underexplored.

2. Methodology

The authors propose a holistic framework involving signal modeling, theoretical bound derivation, hybrid precoder design, and a two-stage localization algorithm.

A. System and Signal Model

Setup: A Multiple-Input Single-Output (MISO) system where a Base Station (BS) equipped with a Uniform Planar Array (UPA) of $M$ Reconfigurable Antennas communicates with a single-antenna UE.
Channel Model: Includes a Line-of-Sight (LoS) path and $I$ Non-Line-of-Sight (NLoS) paths modeled via geometric scatterers.
RA Synthesis Model: Each RA element synthesizes its radiation pattern by combining a finite set of Spherical Harmonic Orthogonal Decomposition (SHOD) basis functions.
- The radiation pattern is controlled by an EM precoder vector ( $e_{m,t}$ ) that weights these basis functions.
- The total channel vector incorporates both the array response and the dynamic radiation patterns of the RAs.

B. Theoretical Analysis (Cramér-Rao Bound)

The authors derive the Fisher Information Matrix (FIM) and the Cramér-Rao Bound (CRB) for the 3D UE position.
The Position Error Bound (PEB) is defined as the trace of the top-left $3\times3$ submatrix of the inverse FIM.
Key Insight: The PEB is shown to be a convex function of the composite covariance matrix formed by the joint digital and EM precoders.

C. Joint Optimization Design

The core contribution is a joint optimization of the digital precoder ( $f_t$ ) and the EM precoder ( $e_{m,t}$ ) to minimize the PEB.

Perfect Knowledge Case: Assuming the UE position is known, the authors formulate a convex optimization problem. They prove that the optimal solution can be decomposed into a low-rank structure, leading to a 4-codeword scheme per transmission block.
Position Uncertainty Case: Since the exact UE position is unknown (otherwise localization wouldn't be needed), the authors design a robust codebook.
- They sample a set of candidate positions within the uncertainty region.
- They formulate a min-max optimization problem to allocate power among codewords such that the worst-case PEB over the uncertainty region is minimized.
- This is solved efficiently as a Convex Semidefinite Program (SDP) using the Schur complement.

D. Localization Algorithm

A two-stage Maximum Likelihood (ML) estimator is proposed:

Coarse Stage: A grid-based search over the uncertainty region to find an initial estimate.
Refinement Stage: A gridless quasi-Newton optimization starting from the coarse estimate to achieve high-precision localization, overcoming grid quantization errors.

3. Key Contributions

Novel Signal Model: Development of a synthesis-based signal model for NF localization using RAs, where each antenna element dynamically forms patterns from a finite set of SHOD basis functions.
Joint Precoder Design: Formulation and solution of a joint optimization problem for digital and EM precoders to minimize the PEB. The solution accounts for UE position uncertainty via a robust power allocation strategy.
Theoretical Bounds: Derivation of analytical expressions for the FIM and CRB specific to RA-assisted NF localization, establishing theoretical performance limits.
Algorithmic Innovation: Proposal of a computationally efficient, two-stage ML localization algorithm (coarse grid search + gridless refinement) tailored for the proposed system.

4. Simulation Results

Extensive simulations were conducted using mmWave parameters (30 GHz carrier, 50x50 UPA).

Accuracy vs. SNR: The proposed method approaches the theoretical PEB (around -5 dB error) at high SNR, significantly outperforming conventional fixed-pattern arrays and heuristic directional beamforming.
Robustness: The RA-based approach demonstrates superior robustness in low-SNR regimes compared to baselines.
Basis Functions: Performance improves monotonically as the number of SHOD basis functions ( $Q$ ) increases. When $Q=1$ , the RA system degrades to a conventional array, validating the model.
Interference: Under strong multipath interference (low Line-of-Sight to Multipath Ratio), the proposed RA system maintains better localization accuracy than conventional arrays, though performance degrades as the LoS component weakens.

5. Significance

Paradigm Shift: This work moves beyond the traditional view of antennas as static hardware, demonstrating that dynamic EM control (via RAs) is a powerful tool for enhancing sensing and localization capabilities.
Near-Field Focus: It specifically targets the near-field regime, which is becoming increasingly relevant for 6G, massive MIMO, and close-range IoT applications where spherical wavefronts dominate.
Practical Implementation: By providing a tractable convex solution for joint precoding and a robust codebook design for uncertain positions, the paper bridges the gap between theoretical EM synthesis and practical deployment in wireless networks.
Future Impact: The findings suggest that integrating RAs into ISAC systems can significantly reduce positioning errors without increasing hardware complexity (antenna count), offering a new path for high-precision 6G localization.